Automatization never produces results that are 100% reliable

Alexander Wiechert, CEO Vexcel Imaging in an interview with Coordinates, at the launch of UltraCam Osprey 4.1 large-format aerial camera

Alexander Wiechert CEO, Vexcel Imaging

Why do you call UltraCam Osprey 4.1 of next generation?

The Osprey 4.1 is represents a completely new UltraCam camera architecture with major new developments in literally all components of the camera system. At Vexcel we innovate constantly. We have launched three camera generations in the past, each with several enhancements. The first generation UltraCam, the UltraCamD, was introduced in 2003. The second generation was launched in 2006 with the UltraCamX and subsequent models such as the UltraCamXp and UltraCamLp. The third generation was released in 2011 with the UltraCam Eagle M1 and has since seen several iterative models, the Eagle M2 and M3, and was also the foundation for new camera designs such as the Osprey M3 or the Condor M1. Now it is again time for a revolutionary jump which brings us to the 4th UltraCam generation with the announcement of the Osprey 4.1, the first model based on this new architecture. This same architecture will be applied to other UltraCammodels, such as the Eagle and Condor, within the next few years. Main aspects of the 4th generation camera architecture are: fully CMOS based cameras, new lenses, new electronics and new camera housing. Collectively, these enhancements result in a new industry standard in efficiency, quality and usability to provide utmost customer benefit.

Despite the ongoing challenges, Vexcel was able to launch Osprey 4.1. How was that made possible?

Well, the development of the Osprey started many, many months ago with research activity that led to development tasks and that now results in a final product. We now are weeks into serial production. So the recent challenges presented by world events affected the whole development cycle for only a short period. We worked closely with our supplier to minimize any effects such as running the camera production in shifts to enable less people in the factory while still maintaining high-volume production. And at Vexcel, we were already very much used to a percentage of the team working remotely, collaborating through video conferencing and so forth, so we already had all the necessary technology in place and employees were already accustomed to using those tools and working in that manner. That enabled us to reorganize the office quickly. The team did a fantastic job here and we were able to maintain a high level of productivity.

How has the prevailing Covid-19 situation affected your work in terms of product development, supply chain etc.?

The impact on the product development was rather minimal, with respect to the software. Generally speaking, software developers are able to work remotely, assuming the technology—such as internet bandwidth and remote connections to the office—are in place. We have retained our development process, our workstreams, and our meeting structure is unchanged though moved to video and teleconferencing. This has worked well with very little additional overhead added. Hardware development, camera support and repair are all different. These teams need the office environment. We implemented measures and procedures to ensure a healthy and safe work place for them. Supply chain was, and continues to be, a challenge. Delivery time for components increased and so we increased our inventory significantly to ensure uninterrupted camera production and for support of installed systems.

Could you please highlight role of Vexcel imageries used by various agencies involved in Covid-19 relief measures?

The Vexcel Data Program (VDP) has two elements: “Gray Sky” for natural disaster mapping and “Blue Sky” for continuous mapping. The high-resolution nadir and oblique imagery of the Osprey, in particular, collected under the Blue Sky initiative was used to help in the implementation and organization of Corona measures in cities. Organizations and government agencies such as FEMA used the imagery as basic input for their planning. The imagery made it possible to decide where to put in place aide or temporary hospitals, which streets needed to be blocked to control traffic or pedestrian flow, how to guide people, etc.

Early this year, there was a strategic alliance announced between Vexcel and Verisk, could you please explain briefly what it is all about?

It is all about combining forces in aerial acquisition and imagery production. Vexcel acquired the Imagery Division of Geomni. Geomni is a Verisk company and, among other things, is collecting imagery for the insurance industry just as Vexcel does for our Vexcel Data Program. The merger eliminates redundant data acquisition, combines production forces and merges the existing imagery libraries. The commitment is to drive rapid innovation across imagery and analytics to enter new markets, create new categories, and better serve our customers. The mission is to become the leading global provider of high-resolution oblique and ortho libraries while continuing as the leading provider of highly efficient and accurate commercial mapping systems.

In 2019, Vexcel launched its Data Program in Europe, how successful has the program been? Do you plan any such programs for other parts of the geography?

The Vexcel Data Program has been incredibly successful in the US over the last few years which has led to early flights across Europe, with an initial focus on Germany, in the latter part of 2019. This was a first test that went very well.

The response from our customer base has been overwhelmingly positive and so we have increased our efforts significantly in 2020. We are now using several UltraCam Condors to collect imagery and we plan to soon complete the whole of Western Europe.

Additionally, the UltraCam Osprey 4.1 has been launched for flights in Europe, so we are now adding high-resolution nadir and oblique imagery to our European library. Besides Europe, we began collections in Australia already last year and continue flying there. New Zealand is on the 2020 list along with several other APAC countries

How do you see the growth of photogrammetry and remote sensing going forward? Please highlight few challenges as well.

Photogrammetry as an underlying methodology is only known by, and only visible to, a very small percentage of people in the world. More people talk are familiar with computer vision or analytics, ML, AI, etc. However, photogrammetry will always play an important role as a basic science that enables other technologies to build on top of for meta data extraction, such as analytics.

So we continue to feel strongly about the need for high-resolution, photogrammetric-grade image capture and related camera systems. We believe current trends of remote sensing towards higher resolutions, shorter turn-around cycles remain intact and that we will see much more software solutions building on top of the basic data.

Challenges could be “cost control” – higher resolution often isn’t cheap to achieve. Additionally, “quality control” could become a growing challenge. The more we use and rely on automatically-derived metadata, such as analytics results, the more we need to ask ourselves how reliable are the results are that we are using to make decisions. Automatization never produces results that are 100% reliable. How do we ensure that we don’t make wrong decisions (or at least minimize the impact of a wrong decision), because we are using information that is incorrect?